laboratory studies on the feeding, bioenergetics, and...

LABORATORY STUDIES ON THE FEEDING,

BIOENERGETICS, AND GROWTH OF FISH1

CHARLES E. WARREN AND GERALD E. DAVIS

Department of Fisheries and Wildlife, Oregon State University, Corvallis

INTRODUCTION

An animal is often studied in the laboratory in order to increase under-standing of its success or failure in nature. Since a species is successfulwithin the range of environmental conditions to which its structural,physiological, and behavioural characteristics permit it to adapt, thesecharacteristics are examined in relation to important combinations offactors present in the natural environment. The biological problemchosen for study will largely determine the particular physiologicaland behavioural responses and environmental factors that will beconsidered in experiments. In studying growth in the laboratory, foodconsumption and changes in body weight are usually measured undervarious conditions, but in addition to accumulation in the body theother important fates of the food consumed should also be determined.The quantity and quality of the ration, the metabolic state of the animal,and the energy demands of its maintenance and behavioural activitieswill largely determine the fates of the food consumed and the growthpossible. Factors in the environment, including food, that influence themetabolic state and the activities of the animal should, then, beconsidered in studies of growth. To increase understanding of thegrowth of an animal in its natural environment, studies of growth musthave a fairly complex conceptual framework. Bioenergetics, as taughtby Brody (1945), and Fry's (1947) exposition of the relation betweenenvironment and animal activity provide us with this framework.

1 A contribution from the Pacific Cooperative Water Pollution and FisheriesResearch Laboratories, Oregon State University. Special Report 23o, OregonAgricultural Experiment Station. This paper is based on the results of researchlargely supported by Research Grants G 10732, GB 467, and GB 4469 from theNational Science Foundation and by Research Grant WP 157 from the FederalWater Pollution Control Administration.

175

176 WARREN & DAVIS

BIOENERGETIC APPROACH TO THE STUDY OF GROWTH

Kleiber (I 96 ia) vividly describes the evolution of bioenergetics fromits 18th century beginnings in the experiments of Priestley, Scheele &Lavoisier, through the development of calorimetry starting about thattime, then through early 20th century arguments concerning Rubner's`specific dynamic effect,' to its modern concern with biochemistry,nutrition, and growth. In a very personal account, Kleiber (1961b)tells us something of the monumental contribution of Brody and of hisown role in the development of this approach to the study of growth,an approach that has contributed greatly to animal husbandry.

Those who are interested in quantifying the feeding and growthrelationships of fish owe much to Ivlev (1939a, b, c, 1945, 1947, 1961a, b)for the approaches he has suggested. Winberg (1956) has broughttogether much of what is known about the bioenergetics and growth offish; and the importance of the contributions of the Russian workersand of Fry (1957) and his students is apparent.

In the following discussion of energy budgets for fish, and in laterdiscussions of growth efficiencies, various categories of losses and usesof food energy are referred to frequently. We have prepared a diagram(Fig. 0 to help clarify our use of terms. The digestible portion of theconsumed food undergoes digestion in the alimentary canal, theremainder being passed out as faeces. The digested materials, hereconsidered assimilated materials (Fig. I), enter the circulatory systemand are distributed about the body. Some portion of the assimilatednitrogenous materials is not metabolized and is lost through the kidneys,gills, and skin. The energy of the remaining assimilated materials hasbeen considered the 'metabolizable energy' by Brody (1945) and Kleiber(196 ia), and Brody has termed their energy or heat content their`physiologic fuel value'. Winberg (1956) uses the expression 'physio-logically useful energy' for the energy of metabolizable materials, butthis is not strictly correct, for not all of the energy of the metabolizablematerials is available for physiological processes and growth. As aresult of deamination, which occurs before nitrogen excretion throughthe kidneys and gills, and other conversion processes, a considerablepart of the energy of metabolizable materials is freed and is not availableto the organism for useful purposes. It is the other part of the energy ofthe metabolizable materials which the organism can use that Brody(1945) and Kleiber (1961a) have called the 'net energy' and that cancorrectly be termed the physiologically useful energy.

Net energy(Physiologically useful

energy)

FEEDING, BIOENERGETICS & GROWTH 177

Energy of food materials( Heat of combustion)

4c

Energy of assimilated

Qu materials

Energy of metabolizable

Energy of nitrogenous materials

materials lost through (Physiologic fuel value)

excretion

Energy of faeces

r

Nonutilized energy freedthrough deamination and

other processes

la.D

(SDA)Processes of digestion.movement and deposition

of food materials

Standard Activitymetabolism

I asL

aR

FIG. 1. Categories of losses and uses of the energy of consumed food materials.

The consumption of food by animals results in an increase in theirrates of oxygen consumption and heat production. This increase isknown as the specific dynamic action (SDA) of the food. In cattle,SDA 'ranges from about 8 per cent of the metabolizable energy at

0.5 maintenance to 38 per cent of the metabolizable energy at full feed'(Brody, 1945). It is generally accepted that SDA is not due to energyutilization for digestion, movement, and deposition of food materialbut represents primarily energy freed and lost through deamination ofproteins to be stored as fats or catabolized; corresponding but smallerlosses occur with fats and carbohydrates (Brody, 1945). Nevertheless,we cannot assume that none of the increase in the rate of oxygenconsumption or heat production after food consumption is due to

178 WARREN & DAVIS

utilization of energy for digesting, moving, or depositing of foodmaterials. Accordingly, we will here consider a small part of thisincrease to result from energy utilization for these purposes (Fig. 0.

In order to represent the losses and utilization of the energy of thefood an animal consumes, Ivlev (1939a, b, c, 1945) proposed the follow-ing equation, which is a modification of one Terroine & Wurmser(1922) used for microorganisms:

Q= Q' + Qr+ Qt+ Qv+ Qwwhere

Q= energy value of food consumed,Q' = energy value of materials laid down as growth,Qr = energy value of faecal and nitrogenous wastes,Qt = energy of primary heat,Qv = energy of external work, andQw = energy of internal work.

The symbols we have used in Fig. I differ from those of Ivlev. By`primary heat,' Ivlev appears to have meant the portion of the energy ofmetabolizable materials that cannot be utilized by the animal. A majorpart of this would certainly be SDA, but Ivlev may have been consideringprimarily heat losses resulting from the inefficiency with which energy-rich materials are utilized through the biochemical processes whichmake activity of an animal possible. He used oxygen consumptiondeterminations to estimate 'internal' and 'external' work, and measuredfood consumption, growth, and the energy value of waste products.He then determined primary heat as the difference between the energyvalue of the food consumed and the sum of the energy values of wasteproducts, oxygen consumption, and growth.

Winberg (1956) objected to the determination of primary heat bythis difference. Fundamentally, Winberg's objection was that theorganism's total heat production resulting from energy utilization aswell as loss comes about through oxygen utilization; and Ivlev, indetermining internal and external work by measurements of oxygenconsumption, had already accounted for all heat production. SDA isusually estimated by measuring differences in the metabolic rates of fedand unfed animals; and the procedures Ivlev (1939c) and Ivlev &Ivleva (1948) used in some of their experiments could lead to estimatesof SDA. This is apparently not recognized by Winberg. We must,however, accept Winberg's objection to Ivlev's (1939a) experiments withsheat-fish (Silurus glanis) larvae, and the same objection appears toapply to his experiments with pike, Esox lucius, (Ivlev, 1939b), though


our understanding of his methods apparently differs from Winberg's(1956). Ivlev did not always carefully describe his methods. Thoughwe may not completely understand Ivlev's objectives and methods or beentirely satisfied with the clarity or usefulness of his equation, his papersfirst drew our attention to the possibilities of bioenergetic studies ofgrowth.

Winberg (1956) himself proposes for estimating food consumptionrates, the use of a 'balanced equation,' which is equivalent to thefollowing:

Qc= i.25(Qr + Q0where

Qc=energy of ration,Qr = energy of metabolism,Qg = energy of weight increase,

and where the metabolizable energy is assumed to be o.8 of the energyof the ration. This equation can be expected to yield reasonablyaccurate estimates of the rates of food consumption of fish in natureonly if the estimates for energy of metabolism are based on laboratoryexperiments with fish expending about as much energy in activity,consuming about as much food, and growing about as fast as those innature. Winberg's (1956) discussion of the experiments of other workersfails to convince one of the value of using laboratory measurements of`routine metabolism' in conjunction with growth measurements toestimate the food consumption of fish. To be useful, the concept ofroutine metabolism needs further definition as to the plane of nutritionand the level of activity of the fish.

Although there may be no entirely adequate way of representing theenergy budget of an animal, Ivlev's equation can be rewritten and hisdefinitions refined. Without clear definitions, such an equation losesmuch of its value, for satisfactory methods of obtaining the necessarydata are then less apparent and interpretations may be morequestionable. We will begin with the following equation:

Qc— Qec= Qg+ Qr (I)

whereQc= energy value of food consumed,Qw = energy value of waste products in faeces, in urine, and lost

through gills and skin,Qg = total change in energy value of materials of body (growth), andQr = energy metabolically utilized or released in all ways for all

purposes.

13

18o WARREN & DAVIS

Now if we let

Qr= Qs+ Qd+ Qa (2)

whereQ8 = energy equivalent to that released in the course of metabolism

of unfed and resting fish (standard metabolism),Qd=-- additional energy released in the course of digestion, assimila-

tion, and storage of materials consumed—SDA, andQa = additional energy released in the course of swimming

activity,

then, substituting, we can represent the over-all energy budget of afish for any given period of time in terms which are convenient for ourpurposes as follows:

Qc— Qw= Qg+ Qs+ Qa+ Qa. (3)

Equation (I) is equivalent to Winberg's balanced equation exceptthat we make no assumption as to the proportion of the consumed foodthat is metabolizable. Davis & Warren (MS, 1967) separated thewaste product term (Q,8) into the energy of faecal material (Q,) andthe energy of materials excreted through the kidneys, gills, and skin(Qv). From a bioenergetic viewpoint, we have defined growth as thetotal change in energy value of the materials of the body. It will some-times be useful to deal with either positive or negative values of thischange. From a physiological viewpoint, growth is often defined as theelaboration of protoplasm, and changes in protein content of the bodymay be considered a more appropriate measurement (Gerking, 1955).Both viewpoints can be usefully combined by substituting a term forprotein changes and a term for fat changes in place of Qg in equations(I) and (3). Use of the bioenergetic viewpoint does not decrease the needfor information on changes in body composition to arrive at an under-standing of the growth process. The elaboration of sex productscan be separated from other changes in body energy value by additionof another term to our equations. The energy metabolically utilized orreleased in all ways for all purposes (Q0 is equivalent to total heatproduction or, with appropriate oxy-calorific conversion, to totaloxygen consumption.

We have attempted to define the components of total metabolism inequation (2) so as to make them independent. Standard metabolism istaken to be the metabolic rate of an unfed fish whose activity has beenprojected to the zero level on a graph of the relationship between itsmetabolic rate and its activity level (Beamish, 1964; Brett, 1964).


Thus it is approximately equivalent to basal metabolic rate. Routinemetabolic rate has been too poorly defined both as to the plane ofnutrition and the level of activity of the animal to be useful. Variationin the behaviour of fish under different conditions makes it difficult toassociate routine metabolic rate with any particular level of activity.Also because of such variation, the measurement of SDA (Qa) in fishcan perhaps best be made by determining the increase in metabolic ratecaused by ingestion of food by a fish required to maintain a fixed lowlevel of swimming activity. From a behavioural point of view, theremay be occasions when it would be useful to partition the energyexpenditure due to swimming activity (Qa) into such components asexpenditures for capture of food, reproductive activity, and randomactivity.

The degree of refinement used in budgeting the energy of an animalwill depend on the objectives of the research. Reasonably accuratebudgeting is only possible under laboratory conditions, but laboratorystudies used in conjunction with studies on the behaviour, food habits,and growth of fish in nature can greatly increase our understanding oftheir bioenergetics and growth.

Basically, two kinds of information can usually be obtained fromlaboratory studies of the bioenergetics of fish: (I) amount and qualityof food consumed, body weight and composition changes, and amountsand kinds of waste products; and (2) amounts of oxygen utilized or ofcarbon dioxide released. We will delay discussing food quality, bodycomposition, and kinds of wastes, all of which have received too littleattention, and consider how information on the heat content or caloricvalue of food, body, and waste materials might be obtained.

Oxygen-bomb calorimetry should be more widely used by thoseinterested in the feeding and growth of fish. The equipment is notexpensive and its use requires relatively little of the experimenter'stime. Dry weight measurements without caloric information on thematerials involved are not satisfactory for bioenergetic studies. Whenthe protein and fat contents of the materials are known, satisfactorycaloric values can be computed using typical conversion factors forprotein and fat. The energy value of faecal waste products can perhapsbest be determined with wet combustion methods, because some ofthese wastes will be in solution in the water in which the fish are held.However, excreted nitrogenous wastes will also be present in solution;and we are aware of no method for discrete measurement of the entireamounts of each of these two kinds of wastes. Ammonia and urea

182 WARREN & DAVIS

are not oxidized under the conditions of the iodate methods describedby Davis & Warren (1965) or Karzinkin & Tarkovskaya (1962),though uric acid and perhaps other nitrogenous wastes are partiallyoxidized. When the objectives of particular experiments warrant theadditional effort, available methods make possible more carefulaccounting of major components of the nitrogenous wastes.

An oxy-calorific coefficient can be used to convert measurements ofthe oxygen consumed by fish to the equivalent amounts, in calories, ofenergy resources oxidized. Although the exact value of the coefficientwill depend on the relative amounts of fat, carbohydrate, and proteinbeing metabolized, it will usually be 3.42 cal/mg (4.89 cal/ml) of oxygenconsumed plus or minus 1.5 per cent (Winberg, 1956; Kleiber, 1961a;Brody, 1945).

If the terms in equation (3) are taken to represent energy values ofmaterials utilized or lost in different ways over some convenient periodof time, say 2-4 weeks, then these values are influenced not only byphysiological and behavioural effects of the environmental factors underconsideration but also by changes in body weight, unless the animal isreceiving only a maintenance ration. There is, of course, considerableinterest in the cumulative effects of physiological, behavioural, andweight changes, but there is also great interest in primary changes inrates of energy utilization or loss caused by changes in environmentalfactors.

In this paper, we will use mean body weight (or energy value) and aunit time of I day to convert energy utilizations and losses occurringduring experimental periods into rate terms. We have not worked withfish of greatly different size, and such a treatment of the data is adequatefor our present purposes. However, there is considerable evidence thatgrowth and metabolism in fish are more nearly proportional to the o.8power of weight than to weight alone (Winberg, 1956; Parker & Larkin,1959; Paloheimo & Dickie, 1966). For the terms in equation (3) to betreated as rates per unit body weight per unit time, for the rates to beadditive, and for the equation to balance, all of the rates must bedetermined as the same function of body size. Strictly, we cannot expectall utilizations and losses of energy to be the same function of body size.Moreover, the value of such equations is that they enable us to compareanimals under different environmental conditions, and changes inenvironmental conditions may change the appropriate function ofbody size. However, it may sometimes be useful to assume all utiliza-tions and losses of energy to be proportional to the same power function


of weight and to use some mean value such as o.8 for this power. Wecould then rewrite equation (3) with rate terms as follows :

Ac—Aw=Ag+As+Ad+Aa (4)

where the Q's of equation (3) and the A's of equation (4) are given by

Qi = Ai Wxt, (5)and

A1= Qt ,

wxtwhere

W= mean body weight or energy value,t = time in days,x= some mean power of W, andi=c, w, g, s, d, or a.

We must distinguish between the structural and physiologicalcapacity of a fish to consume food and the amount of food the fish canand will consume under a given set of environmental conditions.Food capacity may be some simple function of weight, as Kleiber (1933,1961a) has shown it to be for chickens, rabbits, sheep, swine and steers.He found the ratio of food capacity to W°•75 to be reasonably constanteven between species. For some species of fish, standard metabolism(Q8) and total metabolism (Qr) under different conditions of feeding andactivity appear to be proportional to about W' (Winberg, 1956;Paloheimo & Dickie, 1966; Fry, 1957). If this is true, food handling andactivity components of metabolism must also be approximately propor-tional to W0.8 , or be in some way compensatory. Job (1955) showed theactive metabolic rate (Q8 + Qa when Qa is maximal) of brook trout

(Salvelinus fontinalis) to be proportional to Wx, where x=0.75 to 0.94depending on temperature. Brett (1965) has demonstrated that withincreasing activity the powers of weight that give proportionalityincrease from 0.78 with no activity to 0.97 at the maximum sustainedperformance of sockeye salmon (Oncorhynchus nerka). With the aboveconsiderations in mind, perhaps we can assume that Oa and Qd willalso be approximately proportional to W0.8.

The above cited authors discuss some of these relationships ratherthoroughly, and here further consideration need not be given thesematters. Our problem is essentially the practical one of representing ourdata in terms that help to reveal the underlying bioenergetic relation-ships. At this time, too little is known of the physiology, behaviour and

(6)

184 WARREN & DAVIS

ecology of fish to make it possible to represent properly the biologicalinterrelationships in mathematical models of bioenergetics and growth,if, indeed, biological complexity will ever permit such simplerepresentation.

ENVIRONMENT, BIOENERGETICS AND SCOPE FOR GROWTH

Our bioenergetic equation describes an animal only under one set ofenvironmental conditions, and a broader conceptual framework isrequired if the organism—environment complex is to be analyzedeffectively. Those of us interested in the autecology of animals areindebted to Fry (1947) for his statement of the relationship betweenenvironment and activity, 'The idea of activity as distinct frommetabolism and the concept that environment influences activity byacting on metabolism . . . ' ; for his classification of the environment;and for his concept of 'scope for activity' as the difference betweenactive and standard metabolic rates under given environmental condi-tions. Growth can be considered an activity, so perhaps the differencebetween the energy of the food an animal consumes and all other energyutilizations and losses can be defined as the 'scope for growth' underparticular environmental circumstances. Scope for growth, as defined,is not strictly equivalent to scope for activity. Food consumption rate isnot active metabolic rate, but will be influenced by and will influencemetabolic rate. Other metabolic utilizations and losses are not equivalentto standard metabolism. Nevertheless, if we may be permitted thisliberty with Fry's concept, we will be able to take advantage of his wayof relating the performance of an animal (growth in this instance) tothe action of factors in the environment.

Fry classified environmental identities or factors as lethal, masking,directive, controlling (e.g., temperature), limiting (e.g., oxygen), oraccessory, according to the manner in which they may influence anorganism. A particular identity may act under more than one categoryof the classification.

Environmental factors such as temperature and oxygen can influenceeither physiologically or behaviourally the food consumption and themetabolism, and, consequently, the scope for growth of an animal.Food availability and quality are other environmental factors thatmust be considered. They will influence consumption (Q,), wasteproducts (Qu,) and some components of metabolism (Qd and Qa).Food will not influence standard metabolism (Qs) except when its lack


reduces the animal to starvation, when it could be considered a lethalfactor. Perhaps food can be included within Fry's limiting factorcategory. Certainly food acts as a directive factor, and one can envisionit sometimes acting as a masking or an accessory factor.

Fry (1947) has explained how limiting and controlling factors as wellas other possible physiological stressors may act together to determinean animal's scope for activity. The energy budget of an animal providesus with a measure of its scope for growth under particular environ-mental conditions. Together these approaches provide a conceptualframework for examining the relations between growth and environment.

Hari Sethi of our laboratory has permitted us to use his preliminarydata on the bioenergetics and growth of a cichlid (Cichlasoma bimacula-tum) to illustrate the influence of temperature on scope for growth.Different groups of cichlids, which had been acclimated to the testtemperatures of 20, 24, 28, 32, and 36° C while on maintenance rations,were then permitted to feed ad libitum on Tubifex for 14 days. Othergroups were starved for the 14-day period. Weight and caloric deter-minations were made on subgroups at the beginning of this period,and on the remaining test animals at the end. The weights and caloricvalues of the Tubifex consumed were measured by appropriate methods.The percentages of consumed food lost in faeces were determined bythe iodate wet combustion method (Davis & Warren, 1965) for other

TABLE I. Mean percentages of dry weight, range of caloric values and mean caloricvalue in kilocalories per gram dry weight of experimental materials

Number Mean percentMaterial of samples dry weight

Range of Meancaloric values caloric value

Midge larvae 6 14.2 4.97-5.33 5.27

Housefly larvae 8 19.9 6.03-6.16 6. i 1Housefly adults II 26.7 4.76-4.79 4.77Stonefly naiads 19 19.2 4.84-5.48 5.36

Tubificid worms 14 16.2 5.34-5.99 5.49Trout (Salmo clarki)

Low ration 7 18.5 4.14-4.49 4.22

Intermediate ration 9 20.5 4.51-4.72 4.54High ration 12 22.4 4.88-5.17 5.04

Sculpins (Cottusperplexus)

17 23.5 4.74-5.44 5.29

Cichlids (Cichlasomabimaculatum)

130 4.50-5.94 5.15

186 WARREN & DAVIS

groups of fish also fed ad libitum at the various temperatures. The dryweights as percentages of wet weights and average caloric values of foodorganisms and fish used in this and later experiments we will discussare given in Table 1.

Food consumption and scope for growth in calories per fish per dayand assimilation efficiency were greatest at 28° C and decreased at lowerand higher temperatures (Fig. 2 and Table 2). The energy utilization ofthe starved fish does not suggest that standard metabolism (Qs)together with a probably variable cost of activity (Qa) (Fry, 1964)accounted for a very large portion of the energy of the food consumed.

500

450

400aO

350.c

300

4'2500

(..j 200

150

100

50

20 21 22 23 24 25 26 27 28 29 30 31 32Temperature (°C)

Flu. 2. Influence of temperature on the food consumption, losses and uses ofenergy of food materials, and scope for growth of young Cichlasoma bi-maculatum fed ad libitum for a period of 14 days, shown as calories per fishper day. Categories of faecal wastes and specific dynamic action includesmall percentages of nitrogenous wastes. Estimates of metabolism of starvedfish were based on experiments with fish larger than those used for otherestimates and have been adjusted for size differences (Data of Hari Sethi).

Specific dynamic action (Q d) accounted for more energy than did faecallosses and starvation metabolism combined at all temperatures wherethis comparison can be made. Faecal losses represented 3o per cent ofthe energy of the food consumed at 36° C, but starvation metabolismwas not determined at this temperature. The wet combustion method ofdetermining the energy value of faecal wastes may result in the oxidation

33 34 35 36


0ti

07.;

4:5

(:5 ot-- 0 .4- V')

IIIII

O•ct in -

On•• 14

•S'). .s@n

0 0 —10-0

• clen Cn ten 0 l0lg lC)CF O kr, GO en

an CO 0 en 1"--l0 v", 0M en 0 \

CC

'LT1.71 N0 0-0

•E 7!)

•5te,4(1.)CI, ow0 }tgA

,12

•—O 0

O S 71)0 'FA4,4 en 0

Tt '00 01 00 el 0

(4' e;

m- Ctn en 0 l01-- el

∎••0 NO

CP, CO 0 en t--lg lg 01 0en en Ch N

3 1 1 1 1 1

00

0

cct

CC

0, 0ety0

.1zt

17)

,g •g 7:1o0

en I.-- 0 el0 an t•-•Cr, I-- CA 1 1 1 1 1

0wO °00 to,

=r,0 •

.gm'0 tf)

.0 ea

O (.0o

°

g 09 0sa C.)

<4 45.

00 0 n.0 O %.0. . . • .N ',et 00 v", 0%

oo ceo 00 nt0I( 1 1 1

C0E`)

O

fl

°(L)

mv 500

N 0 0 00 000 I-- en 0'0

tsi v,\3' v5 41 1 1 1 1

.cd

U0 .0 (,)0 -0 UN `1-)-0

•—n

ei

›-• 4 ' .0,

tl.)0 0 C.,0 0 0

e‘t 00 ooort-, mo

00 00 co n.c.,s;'s

N Vl N0 N N 0ct.n cc, NIn le le; le

0 00 a.)

Ore,

cd w

F

E G0

0 -4. 00 el l0(.1 fn re)O M lgN el N N en

P. 00 00 0-e)

188 WARREN & DAVIS

of a part but not all of the nitrogenous wastes excreted through thekidneys, gills, and skin. In consequence, small portions of the energyvalues given for faecal wastes and for SDA, which was determined bydifference, properly should be attributed to such nitrogenous wastes.The methods used here do not permit quantitative separation of thiswaste category, but further separation would not greatly alter the aboverelationships since the energy value of the nitrogenous wastes of fish,even on a maintenance ration, is only in the order of 7 per cent of theenergy value of the food they consume (Winberg, 1956).

The effect of an environmental factor on the bioenergetics of ananimal must be considered not only in terms of total energy utilizationand loss but also in rate terms. The cichlids at 36° C consumed morefood per unit of body material than did those at the other temperatures(Fig. 3). Though starvation data are lacking at this temperature, poorutilization of food for growth probably resulted not only from poorassimilation but also from metabolic failure causing increased SDA.This also occurs at 20° C. Kinne (1960) found the total consumption offood by the desert pupfish (Cyprinodon macularius) to be greatlyincreased at high temperatures and efficiency of food utilization forgrowth greatly decreased. There were indications that this was at least

200

180

>.160a

20

1 l t t t

27 28 29 30 31 32 33 34 35 36Temperature (°C)

Fin. 3. Influence of temperature on the food consumption, losses and uses ofenergy of food materials, and scope for growth of young CichIcaoma bi-maculatum shown as calories per mean kilocalorie of fish biomass per day.See legend for Fig. 2 for other information (data of Hari Sethi).

0 (111111(1(

20 21 22 23 24 25 26


in part due to very poor assimilation. Anderson (1959) found thepercentages of consumed protein retained by bluegill (Lepomis macro-chirus) at 8o, 70, 6o, and 5o° F to be 42, 35, 35 and 14, respectively,indicating a considerable increase in deamination and SDA at the lowesttemperature, since protein assimilation was essentially constant.

Sethi has also determined SDA by taking the difference between themetabolic rates of cichlids swimming at a fixed low velocity when notfed and when fed ad libitum at each of the test temperatures. Total SDAfor the food consumed was greatest at 36 and 20° C even though thefish consumed much more food at intermediate temperatures. SDA pergram of food consumed was about four times as great at 36° C and aboutthirteen times as great at 2o° C as at 28° C.

The influence of dissolved oxygen concentration on the ad libitumfood consumption and growth of coho salmon (Oncorhynchus kisutch)fed amphipods (Hermann et al, 1962), of those fed Tubifex (Fisher,1963), and of largemouth bass (Micropterus salmoides) fed earthworms(Stewart, 1962) has been studied in our laboratories. The food consump-tion and the growth of these fish declined with any appreciabledecreases in oxygen concentration below the air saturation level.Marked changes in gross food conversion efficiency occurred only atconcentrations below about 4 mg/l. Moderate decreases in dissolvedoxygen result in reductions in scope for growth apparently by limitingthe amount of food the fish will consume. Fry (1957) has suggested thatthe maximum metabolic rate of fish is restricted by respiratory surfaceand that, in consequence, this 'surface rather than the absorptivesurface of the digestive tract may be the limit of the maximum rate ofgrowth of which fish are capable.' Sethi was able to detect little or nodifference in the metabolic rates of fed and unfed cichlids when thesefish were forced to swim at high velocities. The fish were apparentlyunable to digest or utilize the consumed food when swimming perfor-mance required all of the oxygen they could obtain from theirenvironment.

Ivlev (1961b) has examined food as a factor determining the scope forgrowth of a fish by relating consumption, respiration, and growth tofood density. He determined the growth rate of the plankton-feedingbleak (Alburnus alburnus) at one food density in a hatchery pond.By means of a mathematical model, he then attempted to estimate whatthe consumption, respiration, and growth of this species would be atdifferent food densities. Necessary data on maximum rations androutine and active metabolic rates were obtained from laboratory

190 WARREN & DAVIS

studies. His analysis indicated that consumption rates would increase,respiration rates would decline, and growth rates would increase withincreases in food density. He assumed that changes in respiration wouldbe due to changes in swimming activity necessary to capture food anddid not consider changes in respiration that would result from changes infood consumption rate. For reasons to be discussed below, we believethe latter effect to be the more important, but this does not detractfrom the imaginative manner in which Ivlev related scope for growthto food density.

EFFICIENCIES OF FOOD UTILIZATION FOR

GROWTH AND MAINTENANCE

Discussion of food consumption and growth will lead us to consider theefficiency with which food is utilized for growth or maintenance underdifferent circumstances. There are two general types of efficiencieswith which we will be concerned, total efficiencies and partialefficiencies (Kleiber, 1961a). Each type of efficiency may be employedwith any of the food energy categories, total, digestible, metabolizable,or net (Fig. t). It is necessary to specify the type of efficiency and thefood energy category when discussing efficiencies. Throughout thispaper, the total food energy category will be referred to unless statedotherwise.

Total efficiency, Et, is the gross efficiency of Brody (1945) and Brown(1946, 1957):

Et = G

(7)

whereG = growth, andI= food intake.

Ivlev's (1939a, 1945) coefficient of growth of the first order (10 istotal efficiency on the basis of the total food energy category; hiscoefficient of the second order (K2) is total efficiency on the basis ofmetabolizable food energy.

Partial growth efficiency, Epg, is the net efficiency of Brody (1945)and Brown (1946, 1957):

Epg m (8)

whereM= maintenance ration.


This ratio expresses the efficiency with which an animal utilizes forgrowth that food consumed over and above the amount it would requireto just maintain its tissues. Brown (1946) used the digestible food energycategory in calculating net efficiency.

Partial maintenance efficiency, Epm, is an expression of the efficiencywith which an animal utilizes rations at or below the maintenance rationto maintain its tissues or to prevent them from being catabolized (Kleiber,1961a):

LpEpm =—

ipwhere

Lp=tissue loss prevented, and/p =part of ration preventing loss.

Partial maintenance efficiencies appear not to have been used by fisheryscientists. Considering that many if not most species of freshwater fishin the temperate world grow during only about 6 months of the year,there should be not only theoretical but practical interest in theefficiency with which these fish maintain themselves, avoiding deathfrom starvation or from a combination of starvation and otherenvironmental factors as shown by Ivlev (1961a).

Total efficiencies can be determined with information on growth atone ration. To determine partial growth efficiencies, the maintenanceration also must be known. Partial maintenance efficiency can bedetermined only when data on weight changes corresponding to twodifferent rations at or below the maintenance ration are available.It is the ratio of the difference between tissue lost at the two rations tothe corresponding difference between the two rations.

With increasing ration, gross efficiency increases from zero at themaintenance ration and asymptotically approaches a maximum at themaximum ration if net efficiency is constant. If net efficiency declineswith increasing ration, gross efficiency will increase to a maximum withration increases to some intermediate level and then decline with furtherincreases. Paloheimo & Dickie (1965), after examining literature on thefood consumption and growth of fish, conclude that gross efficiencydecreases from a maximum value at low feeding levels and is indepen-dent of the size of fish. They recognize certain restrictions that must beplaced on this generalization, but our experience leads us to believetheir statement is not sufficiently restrictive. Unless net efficiencydeclines drastically with increases in ration from low levels, and this is

(9)

192 WARREN & DAVIS

not generally the case, gross efficiency will not decline with increasingration except beyond some intermediate level of feeding. Within anygrowth stanza (Parker & Larkin, 1959), the gross efficiency of aparticular fish must decline with increasing size, as a greater portion ofthe food the fish can obtain is necessary for maintenance, and itsgrowth rate will approach zero. If the fish crosses some ecological orphysiological threshold which permits more efficient procurement orutilization of food, its gross efficiency and growth rate may increaseagain. Animals of different races or species and of greatly different sizesmay exhibit nearly the same absolute growth with the same absoluteration (Kleiber, 1926, 1933, 1936), but this is true only when increasesin maintenance costs are compensated by increases in growth rates,which may occur in different races at equivalent physiological ages(Brody, 1945). Faster growing individuals in the same group could havegross efficiencies similar to those of slower growing individuals eventhough the former were larger.

Net efficiencies tend to decline with increasing ration size because ofdeclining assimilation efficiency and increasing SDA, but under someconditions they can be remarkably constant over a wide range ofration sizes, as we will show later. Unless very precise data are available,net efficiencies (partial growth efficiencies) should probably not becomputed for animals that are barely growing, since most of the energyof the food consumed is being used for maintenance requirements.Difficulty in precisely measuring small amounts of growth and thedifferences between maintenance rations and rations little above thesecan lead to unreasonable values for net efficiencies such as those nearand over ioo per cent found by Brown (1946, 1957), unreasonable evenfor the digestible food energy category.

FOOD CONSUMPTION, ACTIVITY,

RESPIRATION, AND GROWTH

In nature fish obtain and utilize food for maintenance and growth undervarious environmental conditions. Not only food availability andquality but other environmental factors determine how much food isavailable for growth and how much is lost or utilized in other ways.Laboratory studies of growth will be most useful if they permit theexamination of the different uses fish make of food when consumingwidely different amounts of foods of different quality under a variety ofconditions.


We have studied the food consumption and growth of individualsculpins (Cottus perplexus) which were held separately in aquaria atseasonal temperatures and fed midge larvae. Some individuals receivedrations near or below the maintenance level, others were fed well abovethis level, and some were given all the food they would consume. Wetweights were determined at the beginning of an experiment, and wet anddry weights at its conclusion. Dry weights of typical individuals were alsodetermined at the beginning. Heats of combustion of samples of fishand of food organisms were used for converting dry weight data tocaloric values (Table f). Percentages of food consumed that wereassimilated were determined by the wet combustion method describedby Davis & Warren (1965), and mean values were used (Table 3).

TABLE 3. Percentages of food consumed that were assimilated bycutthroat trout and sculpins

Percentages offood assimilated

NumberSpecies of determinations Range Mean

Trout (Salmo clarki) 6 84.9-86.1 85.5Sculpins (Cottus perplexus) lo 78.4-84.4 81.9

The use of mean values for assimilation efficiency in this and laterexperiments does not take into account differences in assimilation withdifferences in ration size or season. The means are for fish held at fo° Cduring the winter and fed somewhat less than a maximum ration, andthey may approximate minimum assimilation efficiencies. This willresult in some error in the respiration values, determined by difference,but this error is probably not sufficient under the conditions of theseexperiments to alter appreciably our general findings.

Maximum food consumption and food utilization for growth in a fallexperiment differed strikingly from those observed in a winter experi-ment (Fig. 4). Similar differences have occurred in other experimentsconducted during different seasons (Davis & Warren, 1965; Brocksen,1966). In the fall experiment, net efficiencies declined moderately withincreases of ration size over the wide range of rations the fish wouldconsume; consequently, the maximum gross efficiencies were achievedat two-thirds the maximum ration and declined beyond this Level(Table 4). Energy losses and utilizations in the respiration category

194 WARREN & DAVIS

(Qr) were about four times as great at maximum rations as at rationsnear the maintenance level. This can be attributed almost entirely toincreased SDA (Qd). Fry (1957) states that 'Job (1954) in a singlepreliminary experiment found that the consumption of minnows bytwo yearling S. fontinalis fed to repletion required a rate of oxygenconsumption for the assimilation of this food equivalent to their activeoxygen consumption as determined in the rotating chamber.' Paloheimo& Dickie (1966), making use of Winberg's (1956) balanced equation,found on examining the data of Dawes (193o-31a, b), Pentelow (1939)and Gerking (1955) that the level of metabolism of the experimentalfish increased 4 to 5 times with an increase in ration level from main-tenance to maximum. Not only increases in total daily ration but alsoincreases in individual meal size will increase SDA. The same totalration given in a greater number of feedings will result in a reductionin SDA (Brody, 1945; Cohn, 1963) and in more efficient utilization offood for growth.

Sculpins consumed much less food in the winter experiment than inthe fall experiment. At rations slightly above the maintenance level,their gross efficiencies in winter were about the same as those observedin the fall, but gross efficiencies in winter remained nearly constant with

10

8

-0 6

0C)-Y 4

02.c

-5O

—2

-4 I 1

0-0 Fall, 1961.

0- — Winter, 1962

l I I I I0 5 10

Food15 20 25 30

consumption (cat/kcal/day)35 40

FIG. 4. Relationships between yearling sculpin (Cottus perplexus) food con-sumption rate and growth rate determined in aquarium experiments con-ducted during the fall, 1961, and the winter, 1962 (from Davis & Warren,1965).

0


increases in ration size (Table 5). Net efficiencies at rations somewhatabove the maintenance level were similar in the two experiments, butonly in the winter experiment did slightly higher rations result in markeddecreases in net efficiency (Fig. 5). Respiration accounted for about asmuch energy at consumptions near 1,60o calories in the winter experi-ment as at consumptions near 2,000 calories in the fall experiment(Tables 4 and 5). The decline of net efficiencies over such a narrowrange of rations was probably primarily due to increases in SDA.

60-o--o Fall, 1961

0

q-- -a Winter, 1962

20 I I I I I I

0 5 10 15 20 25 30 35Food consumption (col /kcal/day)

FIG. 5. Relationships between food consumption rate and net efficiency offood utilization for growth of yearling sculpins (Coitus perplexus) in aquariumexperiments conducted during the fall, 1961 and the winter, 1962 (data fromDavis & Warren, 1965).

Such increases can be expected to occur when animals are unableeffectively to utilize food for growth either because of nutrientimbalances (Brody, 1945) or because of the metabolic state of the animal.Under some circumstances, metabolic processes may limit food utiliza-tion for growth to rates below those possible under other circumstances.If an animal is unable to synthesize protein from the amino acidsavailable in its food, deamination will increase SDA. Sculpins andother fish may be metabolically unable to utilize food for growtheffectively during periods of the year when they do not normally grow,though they must be able to obtain and utilize enough food to maintainthemselves during these periods. In the winter experiment, partialmaintenance efficiencies also declined with increases in ration size;at the lowest rations these efficiencies approximated the highest net

14

WARREN & DAVIS196

O ;"8w ryrj

WI O V 00 N 1/40 01 tn o et oo• • • • • • • • • • • •h0 .1- et en .1- oo h o0 1/40.1- kr) kr) h .1- en .1- .1- et cn en en

00 ON en 0 1/40 VD .1- M10 7N MN N 0-I- oo N 00n••n N enN enenenenN enen

M 00 N cn 0 N 1/40 op InN 0 .1- 00 Q M en 00 0 r•-• NM cn en et '.O h 00 1"-- 00 0 0.,

I. .1

0 40 N N N N V 00 N10 0 N N 0 0 ON t-- nO

N en V) 1/40 1/40 kr, 10 oo

.1- G7N 0 00 N k.0ooen ONN en d. 0 1m et .1- 00 oo oo 1: 1- 1/40 r

1.4 w (.4‘

N a1 N et n.0 n-n

00 01 01 C.- 00 00 Nenenenenen

trl v.) .1- .- L/1 0 h °°‘.0 0 a N kr) 0.n 0 N n••n.1. kr) horN 0 0 N. I-7„ 00 0 ‘,0

o-r N N

1/40 1/40 N h. 0 41 N.1- .1- tr) .1- 41 00 CP, ON 0 VI17 c?, '1"!

11 N en .1- kr110 on-n

4§t44 0.0


N.o100041v6 od w; nC;

ct.etNmNN

•cr N 0 o et oe;knu1111111knen'td-Men

O INO m 4-0.0

41 N et on V:;) 00 on N 'eh CO t-on on et 00 on UP 00 un on on un CTm on on on et ..eo on Up VP CT C) C)

bo .0 00L" 0

bhoN5

50 o

OVICON

t0r-..111.000Nco

O-Ot- ontrmt-.00.et-1 7 1

, N N N

Cel i•u.z_c-n bfe

"0 6o o

' ,r)

vsvs

0upN

onupN

N eno, 000 en

CSnce,on

O- 0 N Non 00 Net t- N 00

NNN

UP VP0 t-on en

• le)

I—t O4.)

—0 rdga0 0 c

5 §-49

,

4.°

0

^fn 00 0 7 t, N on N 0 GO t- N41 on 040 00 00 On t- co VD 00 0

N on

NeeNt-1/40000MW.Nen000e.,118MCAS

Oett--.WNV10000000VI 0 c:1) 00 '1- n0 r1 0 0

" et 1: 1: 1, ‘4:'

N te, et kfl VD N. 00 01 0

o---o Spring, 1963

•--• Winter, 1964

198 WARREN & DAVIS

efficiencies occurring at rations slightly above the maintenance ration(Table 5). Anderson (1959) found bluegill and Brown (1946) foundbrown trout (Salmo trutta) to have reduced rates of food consumptionand growth during late summer, fall, and winter even though thesefish were held at constant temperatures. These workers suggest thatseasonal changes in growth may be hormonally controlled.

Figure 6 (unpublished data of Brocksen) suggests that cutthroattrout (Salmo clarki) may utilize food for growth under some conditionsin late winter with higher gross efficiency than in the spring, at leastover the narrow range of rations the fish will consume in the winter.

12

9

;'," 6-0

3

0

o —3

—6

—9 0 10 20 30 40 50

Food consumption (cal/kcal/day)

FIG. 6. Relationships between food consumption rate and growth rate ofyearling cutthroat trout (Salmo clarki) determined in aquarium experimentsconducted during the spring, 1963, and the winter, 1964 (data of RobertBrocksen).

The higher gross efficiencies in winter were primarily due to the lowermaintenance costs, probably resulting from lower temperatures. Themaximum ration at this time was about equal to the maintenanceration in the spring experiment. The shapes of the two curves indicatethat net efficiencies in the winter declined very little but in the springthey declined considerably with increases in ration size. These datasuggest that there are seasons and temperatures at which maintenancecosts may be low but at which the fish are still metabolically able toutilize food effectively for growth. Under these circumstances, the grossefficiencies will be high. The trout were fed Tubifex and the experiments

60


performed in much the same way as described above for the sculpins.Experiments like the ones we have discussed were conducted to make

it possible to estimate the food consumption of fish in laboratorystream communities on the basis of their growth. The validity of anyestimates of food consumption based on growth in the laboratorystreams depends on whether or not the energy expenditures forswimming activity (Qa) in the aquaria were similar to such expendituresin the streams. Brocksen (1966) compared the food consumption andgrowth of trout in aquaria to those of trout held at two water velocitiesin laboratory streams in which no food was available other thanhousefly adults and larvae which he intermittently introduced into thecurrent. Water velocities of 24 and 39 cm/s were provided in differentstreams. The trout occupied positions near large rocks and moved intofaster current usually only in search of food. Some of the stream fishand some of the aquarium fish were starved, while others were fed arange of rations extending to the maximum the fish would consume.

The trout in the streams had higher gross efficiencies of food utiliza-tion for growth than did those in the aquaria, but there was no significantdifference between the efficiencies of fish in the streams having differentwater velocities (Fig. 7). The maintenance costs of the fish in the streams

I I i I i I i I I I4 8 12 16 20 24 28 32 36 40

Food consumption (cal/kcal /day)

FIG. 7. Relationships between food consumption rate and growth rate ofunderyearling cutthroat trout (Salmo clarkOheld in aquaria and in laboratorystreams in an experiment conducted during winter, 1966 (data of RobertBrocksen).

200 WARREN & DAVIS

appear to have been less than those of the fish in the aquaria. Observa-tion of the fish in the streams and in the aquaria suggested that therandom activity of the fish in the aquaria was as great or greater thanthe activity of the fish in the streams. Even the starved fish in the streamsentered the faster current, apparently in search of food.

We can estimate the SDA (Qa) at different consumption rates in thelaboratory streams (Table 6) by assuming that other metabolic costswere the same in fish receiving food as in the starved individuals. SDAincreased markedly as the maximum ration was approached and becamemore than double all other respiratory costs, resulting in a decline innet efficiencies. Thus, gross efficiency reached a maximum at slightlymore than two-thirds the maximum ration and then declined.

Allen (1951) and Horton (1961) have used Pentelow's (1939) data onthe food consumption and growth of brown trout in the laboratory toestimate the food consumption of trout in streams. Gerking (1962) usedsimilar laboratory data on bluegill (Gerking, 1955) to estimate the foodconsumption of this species in a lake. Hatanaka & Takahashi (1960),on the basis of laboratory experiments on growth, estimated the foodconsumption of Pacific mackerel (Pneumatophorus japonicus) in nature;and in a series of papers they and their co-workers estimated the foodconsumption of other marine fish by the same method. Warren et al(1964) estimated the food consumption of cutthroat trout on the basisof food consumption-growth relationships established in laboratoryexperiments conducted during the growth season under naturaltemperature and light conditions.

Allen and Horton both assumed that net efficiency of food utilizationfor growth was constant over a wide range of ration sizes, thoughanalysis of Pentelow's data does not suggest net efficiencies to have beenconstant (Pentelow, 1939; Horton, 1961). As shown above, not onlygross but also net efficiencies can vary widely under different conditions.If laboratory studies on the food consumption and growth of fish areto be used for estimating food consumption in nature, the experimentsshould be conducted during the season for which the estimates are tobe made and they should duplicate natural conditions as nearly aspossible with regard to type of food, temperature, and light. It seemsunlikely that swimming activity for feeding accounts for a large part ofthe total metabolic rate (Qr) of growing fish, if one considers that theenergy utilization and loss through SDA (Qa) in feeding and growingfish is relatively large, and that the energetic costs of moderate levelsof swimming activity (Qa) are relatively low (Brett, 1964; Brett &

0

0O


;IS

4.)

.°3E4- 0

0

Q

0

O

0 4-9

O7*;

0 en ON 'I* h N C*1 h 4-1

M v-) 0n0 0‘10

R

V)

0

N N 00 el N 00 .5Cn-n 4'1 CO 0 0 01 en N N•:1- M 'TY N 00 00 00 0

000

0 xiC4:1

i-1

.0

ocul

g 00

cu rz> o

7.1 •2'4)) rd

A-.g 5'CI' %.0 00

.'b

A 5E0-I

00

,

N

.E 0NNNNNNNNN

en en en en en en en en en.1- "et <I- •el-

N h. ,Z) N C V) ON h

en N. a Q1 on 01/40 0

00 tn 0 .4 \rr n-n n- r n-n

CO 00 00 M<00 CiN en .1: CO Cln

- N N

0 CI \ N \ 00 0N M as NN N en en en 'ter

0 0 ":1. el Is- . N%.0 •d• •ch en •el• -1- ON01 r h oo

N N N c,i‘

0 I-- n0 WI h00 0 0QN N •el- co 10 c.") 1.0h 0 VI N ∎C).,0..N,M 4 en en en v") V"eh 4

N 00 0 CT kr)

vz; v;kr) 'I-

00 - 0 ON 0 0I r: 1/46 N. rnmenen

202 WARREN & DAVIS

Sutherland, 1965; Brocksen, 1966). Brocksen's (1966) experiment tendsto support this viewpoint, which leads to the conclusion that perhapslaboratory studies on food consumption and growth can providereliable estimates of food consumption in nature.

Mann (1965) has used Winberg's (1956) balanced equation forestimating food consumption by fish. He assumed the metabolic rate(Qr) of fish in nature to be twice the routine metabolic rate, as Winbergsuggested. When one considers how dependent the metabolic rate offish is on the SDA (Q d), and how much SDA is influenced by foodconsumption rate and environmental conditions, there remains doubtas to how reliable estimates of the food consumption of fish in naturemay be when based on growth and on routine metabolic rate multipliedby any constant.

STOCK BIOMASS, FOOD CONSUMPTION,

GROWTH, AND PRODUCTION

Ivlev (1947) describes the relationship between stock density, growthand production as one in which growth rate declines with increasingdensity, while production increases to some intermediate density andthen declines with further density increases. Beverton & Holt (1957)have also examined some of the possible relationships between stockdensity, food supply, food consumption, and growth, for use in theirtheoretical models of exploited fish populations.

We have investigated the relationships between food consumptionand growth of fish in aquaria for the purpose of estimating foodconsumption on the basis of growth of fish in laboratory streamcommunities (Davis & Warren, 1965; Brocksen, 1966) and in anexperimental trout stream, Berry Creek (Warren et al, 1964). Theaquarium experiments have been performed concurrently with thestream experiments or during appropriate seasons in different yearswith fish of similar ages and sizes and with temperature and lightconditions varying seasonally and daily as in the streams. Meangrowth rates found in the streams have been used to estimate meanconsumption rates from curves like those presented earlier in this paper.

Communities of algae and herbivorous midge larvae were allowed todevelop in 6 laboratory streams which were then stocked with differentamounts and combinations of herbivorous snails, carnivorous stoneflynaiads, sculpins, and trout. The midge larvae produced in the streamswere the principal food resource of the carnivores. The growth and

0...........

0

2.5- A

,..7 2.0- a 0E

-.....„. 1-5 ---0.o..le 1.0 o

\= 0-

•C 0 5 — \

.f.; \

A x

-oodi: — 0 5- o—o Production

6

50-°

79 3

5 2

1.c

0

E

12

10

8 S

6 g0

u_


production rates of the stocked animals have been estimated directly byrecovering and weighing the animals.

Sculpin growth rate declines as their biomass increases (Fig. 8),unless the biomasses of other carnivores are sufficient to alter thisrelationship. Food consumption rate has also been found to declinewith any considerable increase in biomass. Total food consumptionincreases with increasing sculpin biomass to a level at which depletionof the food resource results in a decline in total consumption. Produc-tion, being a function of both growth rate and biomass, increases withincreasing biomass to a point at which the decline in food resourceand increased utilization of food for maintaining the sculpins reducethe growth rate sufficiently for production to decline. In an experimentin which both sculpins and stonefly naiads were used (Table 7), thesculpins lost weight when stoneflies were present at both high and lowsculpin biomasses. Gross efficiencies of food utilization for growth werelow at high sculpin biomasses and were high at low sculpin biomasseswhen stoneflies were absent.

Brocksen (1966) showed that growth rates of cutthroat trout in

A- -A rate \Growth-10-

0••••0 Food consumption NA

-2-5o 1

2 3 4 5 6 7Biomass ( kcat /m2)

FIG. 8. Relationships between the biomass of yearling sculpins (Cottus

perplexus) and their food consumption, growth, and production in a labora-tory stream experiment conducted during the spring, 1961 (from Davis &Warren, 1965).

Z Q °

,0 8 2g to

eN44

O

7:10o •,^7,

cC

7:1

00 0w 2

0U

0 co atg s

E

2 0 .5r0

204 WARREN & DAVIS

Sao

Ot

tr.>• .0

E4) 0

v)000

503

u 0,

n•••n" ,••••n

0rz),

0 •

O 00 gL)

1.1 CI)5

0ocv,v.);-n

›,

UO M-.0 00

0 00en 41 4,"u DO0 •4 .3 00

5 on-n

tet 800c3 ai

.1el) 0

0 a0cs,en a.) g• O

r..-.;O >.

0• 0 Z

°Q 00CU •u 8

•g ,z)W0

Ri_ .

1.4M 8

WI WI e.e. 0 00,4 •te) k

mr) tr)

V.")110 en

N- 0.-.0001 OR N en

n-r

O 0 0 ON 0CO 0 t•-•M .1. 1/4.0In

1O en en N N 000N Nco •••n 10

4 •–, c‘i

N N 00 CO 0N al .+1'Cr CO

000 0 0 0 0oo col ooo VI

• cl" N (4') ri"

O 0en17 I I1/40 n,0^

O 0 0 0 0 0en a, 0 1/41) 0 NN •el- en Q u-1t N 4 N 00"

N en 'et

o—o ConsumptionA— —A Growth rateo—o Production

6 8 10Biomass (kcal /m2)

710

2

02


laboratory stream communities declined at high trout biomasses(Fig. 9), but that total food consumption increased to a high level andremained high with increasing biomass. Thus, the decline in productionfrom the maximum in the case of the trout was due to increased main-tenance costs and not to decline in the food resource. The trout in thelaboratory streams feed almost entirely on drifting midge larvae, pupae,and adults and do not appear to reduce greatly the benthic populationsproducing this drift, as do the sculpins.

12—24

FIG. 9. Relationships between the biomass of yearling cutthroat trout (Salmo

clarki) and their food consumption, growth, and production in a laboratorystream experiment conducted during the winter, 1964 (From Brocksen, 1966).

Data on the relationships between carnivore biomass, food consump-tion, and growth are difficult to obtain in nature, but some of the BerryCreek data (Table 8) suggest that the relationships are similar to thosefound in the laboratory streams. Berry Creek does not support suchhigh trout biomasses as do the laboratory streams, probably becausethe food organisms in Berry Creek are utilized by many other carniv-orous species. The lowest trout biomasses present in two unenrichedsections of Berry Creek during 1961, 1962, and 1963 are nearly optimalfor production, higher biomasses representing overstocking. Foodconsumption did not change greatly with increasing biomass, butgrowth rates declined because of increased utilization of food for

206 WARREN & DAVIS

rn 0 \ •crI ,f) IN

IN 00 N 00

tn.

V100 0 .0 00N- 01 010-00 00 .11M V1 N .oz: n V1 .") vs 1/46

O 0 0 0 0 0

07\ 00 \N VI 01

7:$00 „,

0g• 4.

-ooraO

44 CI

trn 00 0 oovzs 00 co -or0 on or,N00 1/46 tz:

V1N o (MN Nr•-.. 00 N.N 0 000 r(1,

417:)0 5 0 0 0 0 00 V

00 ,0.ss

tr) ‘0. 0 1/40o 0 en

44 00 co °orC.)

5O

0 0 0 0 0 0MOO N N1/40 0 t•-• N o

Cvt; (4; vr ot;

Nen.NM,0 VD 1/40 n.o n4;)aN a crs a% 01

0

NCo •08v)


maintaining the stock. At the highest biomasses, the fish lost weight;gross efficiency of food utilization for growth increased with decreasingbiomass.

Biomass level, then, may indicate the extent to which a species willutilize its food resource. The relationships between biomass and foodconsumption and growth rates may be of use in evaluating the level ofintraspecific competition for food. However, these relationships cannotbe expected to be the same if there are differences in food production orin the level of interspecific competition. Thus, the biomass of a speciesgenerally does not provide us with a measure of the opportunity anindividual of that species has to obtain food.

FOOD DENSITY AND SCOPE FOR GROWTH

The idea that fish do not always have unlimited supplies of food innature is implicit in experiments in which fish are fed restricted rations.The growth occurring at different rations in the laboratory is studied toobtain some insight into how growth might be influenced by foodavailability in nature. Unfortunately, it is difficult to relate rationlevels in the laboratory to food availability in nature, there being noentirely satisfactory solution to the problem of determining the latter.Information on the production of prey species would be of some value,but because of the many different fates of prey, even this informationleaves the question of availability unresolved.

One can reasonably conclude that it is the density of food in ananimal's immediate surroundings that determines the amount of foodthe animal can obtain in a short period of time and the energy cost ofobtaining this food. Density of a prey species is the outcome of its rateof production and the rates at which it is being consumed, being decom-posed, and emigrating. Thus, changes in production, consumption,decomposition, or emigration should be reflected in changes in thedensity of the prey. Mean prey density over a period of time sufficientlylong for measurable weight changes to occur in individual predatorswould appear to be one useful measure of food as an environmentalfactor influencing the scope for growth of the predator. Ivlev (196 ia)has summarized his research on the effects of prey density as well asprey distribution and predator selectivity on the intensity of feeding offish. Brocksen (1966) has demonstrated that sculpins and stoneflynaiads influence their own food density and that of trout in laboratorystreams, but that trout influence their own food density but not that of

208 WARREN & DAVIS

the sculpins and stoneflies, apparently because of differences in feedingbehaviour.

The food density which must be considered is the density which isappropriate in terms of the structure and feeding behaviour of the fishand the life cycles, structure, and behaviour of their food organisms.The appropriate food density for one life history stage or age group offish may not be appropriate for another. Seasonal changes in feedingbehaviour may necessitate the selection of different groups of foodorganisms during different seasons for density measurements if thesemeasurements are to be appropriate. Measurements of the total biomassof benthic or planktonic organisms would not always be appropriate.It would depend on what the fish were eating.

Borutsky (1960) distinguishes between the forage resource, 'the totalcomplex of autochthonous and allochthonous animal and vegetableorganisms and their decomposition products present,' and the foragebase, 'that part of the forage resources utilizable by the existent fishpopulation.' He uses fish population in this context to be all the agegroups of all the species present. By restricting the meaning ofappropriate food density to the density of food available for a particularlife history stage or age group of a particular species, we use this termin a more restricted sense than Borutsky uses forage base.

The absence of an air bladder in the sculpin limits it to a benthicexistence, and its mouth is well adapted for removing food organismsfrom the substrate. In a spring experiment, the food consumption byyearling sculpins appears to have increased almost linearly withincreases in the mean density of their food organisms in the benthos oflaboratory streams (Fig. to). The low food density at which the sculpinscan just maintain their tissues is of considerable biological interest.It perhaps approximates the density at which sculpins could not besuccessful if it were the result of low production of food or intensiveinterspecific competition. Were this low density the result of intensiveintraspecific competition, a decline in the stock would presumablyoccur.

The air bladder, mouth, and behaviour of trout adapt them for feedingin streams on drifting organisms. The extent to which trout removefood organisms from the substrate in natural streams is not known, butin the laboratory streams they feed almost entirely on organisms driftingin the current. In a winter experiment, Brocksen (1966) found thatcutthroat trout in laboratory streams required a density of driftingfood organisms slightly above i cal/ms to obtain a maintenance ration


and lost considerable weight at a lower density (Fig. I O. Consumptionand growth rates increased most rapidly at densities near I cal/m3but continued to increase to the highest density of about to cal/m3.

40

/VV

00('7 \pi'c't`c?il

2 15

20— v,/cose 't Cat

- ,ii

4' • 5

/ i.e6*Alec • • •ites9‘

10cr

5

0

col's

i I i I I0 . 6 0.7 0 8 09 1.0 11

Food density (kcal/m2)

FIG. to. Relationships between density of food organisms in the benthos andthe food consumption, assimilation, respiration, and scope for growth ofyearling sculpins (Cottus perplexus) determined in a laboratory stream experi-ment conducted during spring, 1961.

Determining the appropriate food density to measure in naturerequires in most instances considerably more knowledge than we haveof the feeding behaviour of fish. Detailed food habit information on thespecies and age groups under consideration is necessary, but thepossibility of obtaining this has been demonstrated by many workers(Allen, 1942 ; Neill, 1938 ; Maitland, 1965). The growth rates of cutthroattrout in Berry Creek tend to be highest during periods when thedensities not only of food organisms drifting in the current but of thosein the benthic environment are highest, but the data are too variableto define clearly the relationships. Density of drifting food organismsmust be some function of their density in the benthic environment, sorelationships between appropriate benthos densities and the foodconsumption and growth of trout might be expected regardless of thefeeding behaviour of the fish.

35

30

073' 25

210 WARREN & DAVIS

50—

40—

30

20

0

10

2I I I I—100

0 3 4 5 6 7 8 9 11Food density (cal/m3)

FIG. i 1. Relationships between density of drifting food organisms and thefood consumption, assimilation, respiration, and scope for growth of yearlingcutthroat trout (Salmo clarki) determined in a laboratory stream experimentconducted during winter, 1964 (data from Brocksen, 1966).

It is difficult to find data in the literature which make possibleexamination of the relationships between food density and growth innature. Most studies apparently have not been directed to such an end.Data of Allen (1951) suggest that a relationship existed between thedensity of food organisms in the benthos and the growth rate of browntrout, and data of Horton (1961) indicate a similar relationship betweenbenthos density and total food consumption by brown trout (Fig. 12).Johnson (1961) has shown that, in a series of seven basins in the Babineand Nilkitkwa lake system in British Columbia, the mean weight ofjuvenile sockeye salmon in mid-October exhibits a high positive correla-tion with the mean dry weight of zooplankton per cubic meter frommid-June to mid-October, the period of the year when nearly all of thegrowth of age-group o occurs. The dry weight of zooplankton duringthis period exhibited a high negative correlation with the late Augustdensity of the fish in numbers. In a later paper, Johnson (1965), usinghis own data and the data of other workers from three other lakes,demonstrated that the mean weights of juvenile sockeye in mid-Octoberwhen plotted against the mean density of zooplankton from mid-Juneto mid-October lie remarkably near the same line. The points from each


lake tend to be grouped, the size of the fish being smallest in the lakeshaving low plankton densities and largest in lakes having high densities.This suggests that the growth rate the sockeye can maintain at a givenplankton density is more a characteristic of this salmon than of theparticular lake ecosystem

115-

0--:-..-x 110

.0

6- 105

1000

Hortono I 1957

q I 1958 /

/IE 1957

0//III 1957

/ 0 71 1958o /

III 1958

— II M,R

ov- o

0 NI

AllenI

-

-40

-10

rn

E

300

mi20 2

I.L.

2 3 4 5 6Biomass of bottom fauna (g/m2)

7 8

FIG. 12. Relationships between biomass of bottom fauna and growth rate of

brown trout (Salmo trutta) in the Horokiwi Stream (from Table 66, p. 188,Allen, 1951) and food consumption of brown trout in a Dartmoor stream

(from Table 12, p. 333, Horton, 1961).

Since publication of Lindeman's (1942) paper on the trophic dynamicaspect of ecology, the increased interest in the dynamics of production,consumption, and decomposition has resulted in questions of thebiological importance of densities of animals being given little attention,except perhaps by the Russians (Ivlev, 1961 a, b; Borutsky, 1960).Acknowledging the ultimate importance of these dynamic processes,we must also recognize that we cannot usually hope to measure therates of production and utilization of more than a few species of preyorganisms. Density of prey organisms is an outcome of these dynamicprocesses, an outcome of considerable importance in determining thescope for growth of the feeding fish.

15

212 WARREN & DAVIS

ACKNOWLEDGMENTS

We wish to express our appreciation to Hari Sethi for permission to usehis unpublished data and to Robert W. Brocksen whose research with uson laboratory stream communities has led to some of the ideas wepresent here. For other ideas, we are indebted to our associates PeterDoudoroff and Hugo M. Krueger. These two and James D. Hall werevery helpful in their critical reviews of the manuscript.

REFERENCES

ALLEN K. R. (1942) Comparison of bottom faunas as sources of available fish food.Trans. Am. Fish. Soc., 71, 275-283.

ALLEN K. R. (1951) The Horokiwi Stream: a study of a trout population. Fish. Bull.N.Z 10, 1-238.

ANDERSON R. O. (1959) The Influence of Season and Temperature on Growth of theBluegill, Lepomis macrochirus Rafinesque. Ph.D. Thesis. University of Michigan,Ann Arbor, 133 pp.

BEAMISH F. W.H. (1964) Respiration of fishes with special emphasis on standardoxygen consumption. II. Influence of weight and temperature on respiration ofseveral species. Can. J. Zool., 42, 177-188.

BEVERTON R. J. H. & HOLT S. J. (1957) On the dynamics of exploited fish populations.Fishery Invest., Lond., Ser. 2, 19, 1-533.

BORUTSKY E. V. (1960) (The fishery forage base.) Akademiya Nauk SSSR. Trudy Inst.Morf. Zhivot., No. 13, 5-61. (OTS 63-11129.)

BRETT J. R. (1964) The respiratory metabolism and swimming performance of youngsockeye salmon. J. Fish. Res. Bd Can., 21,1183-1226.

BRErr J.R. (1965) The relation of size to rate of oxygen consumption and sustainedswimming speed of sockeye salmon (Oncorhynchus nerka). J. Fish. Res. Bd Can.,22, 1491-1501.

BRETT J.R. & SUTHERLAND D.B. (1965) Respiratory metabolism of pumpkinseed(Lepomis gibbosus) in relation to swimming speed. J. Fish. Res. Bd Can., 22,405-409.

BROCKSEN R.W. (1966) Influence of Competition on the Food Consumption andProduction of Animals in Laboratory Stream Communities. M.S. Thesis, OregonState Univ., Corvallis, 82 pp.

BRODY S. (1945) Bioenergetics and Growth. New York, Reinhold. 1,023 pp.BROWN M. E. (1946) The growth of brown trout (Salmo trutta Linn.) II. The growth of

two-year-old trout at a constant temperature of 11.5° C. J. exp. Biol., 22, 130-144.BROWN M.E. (1957) Experimental studies on growth, pp. 361-400. In Brown M.E.

(Ed.) The Physiology of Fishes, Vol. 1. New York, Academic Press, Inc.COHN C. (1963) Feeding frequency and body composition. Ann. N.Y. Acad. Sci.,

110, 395-409.DAVIS G.E. & WARREN C.E. (1965) Trophic relations of a sculpin in laboratory

stream communities. J. Wildl. Mgmt., 29, 846-871.


DAVIS G. E. & WARREN C.E. (1967 'MS) Estimation of food consumption rates.

In Ricker W. E. (Ed.) IBP Handbook on Freshwater Fish Production. In Press.

DAWES B. (193o-31a) Growth and maintenance in the plaice (Pleuronectes platessaL.) I. J. mar. biol. Ass. U.K., 17, 103-174.

DAWES B. (,93o-31b) Growth and maintenance in the plaice (P. platessa L.) II.

J. mar. biol. Ass. U.K., 17, 877-975.FISHER R. J. (1963) Influence of Oxygen Concentration and of its Diurnal Fluctuations

on the Growth of Juvenile Coho Salmon. M.S. Thesis, Oregon State University,

Corvallis, 48 pp.FRY F.E. (1947) Effects of the environment on animal activity. Univ. Toronto

Stud. biol. Ser., 55 Publ. Ontario Res. Lab. No. 68, pp. 5-62.

FRY F.E.J. (1957) The aquatic respiration of fish, pp. 1-63. In Brown M.E. (Ed.)

The Physiology of Fishes, Vol. i . New York, Academic Press, Inc.FRY F.E.J. (1964) Animals in aquatic environments: fishes, pp. 715-728. In Dill

D.B., Adolph E.F. & Wilbur C.G. (Eds.) Adaption to Environment, Section 4,

Handbook of Physiology. Washington, D.C., Am. Physiol. Soc.GERKING S.D. (1955) Influence of rate of feeding on body composition and protein

metabolism of bluegill sunfish. Physiol. ZoOL, 28, 267-282.

GERKING S.D. (1962) Production and food utilization in a population of bluegillsunfish. Ecol. Monogr., 32, 31-78.

HATANAKA M. & TAKAHASHI M. (1960) Studies on the amounts of the anchovy

consumed by the mackerel. Tohoku J. agric. Res., II, 83-10o.HERMANN R.B., WARREN C.E. & DOLTDOROFF P. (1962) Influence of oxygen con-

centration on the growth of juvenile coho salmon. Trans. Am. Fish. Soc., 91,

155-167.HORTON P.A. (1961) The bionomics of brown trout in a Dartmoor stream. J. Anim.

Ecol., 30, 311-338.

IVLEV V.S. (1939a) The energy balance of the growing larva of Silurus glanis.C. r. Acad. Sci. Moscow, N. S., 25,87-89.

IVLEV V.S. (1939b) The effect of starvation on energy transformation during thegrowth of fish. C. r. Acad. Sci. Moscow, N. S., 25, 90-92.

IVLEV V.S. (1939c) Energy balance in the carp. Zool. Zh., 18, 303-318.

IVLEV V.S. ( 1 945) [The biological productivity of waters] Usp. sovrem. Biol., 19,

98-120. (Fish. Res. Bd Can. Transl. Ser. No. 394.)IvLEv V.S. (1947) [Effect of density of planting on the growth of carp] Byull. mosk.

Obshch. Ispyt. Prir., 52, 29-38.

IVLEV V.S. (1961a) Experimental Ecology of the Feeding of Fishes. New Haven,

Yale University Press. 302 pp.IVLEV V.S. (1961b) [On the utilization of food by plankton-eating fishes]. Trudy"

Sevastopol' biol. Sta., 14, 188-201. (Fish. Res. Bd Can. Transl. Ser. 447.)IVLEV V.S. & IVLEVA I.V. (1948) Conversion of energy in the growth of birds.

Byull. mosk. Obshch. Prir., 53, 23-37.Jos S. V. 0955) The oxygen consumption of Salvelinus fontinalis. Univ. Toronto Stud.

biol. Ser., 61, 1-39.JOHNSON W.E. (1961) Aspects of the ecology of a pelagic, zooplanktoneating fish.

Verh. int. Verein. theor. angew. Limnol., 14, 727-731.JOHNSON W.E. (1965) On mechanisms of self-regulation of population abundance in

Oncorhynchus nerka. Mitt. int. Verein. theor. angew. Limnol., 13, 66-87.

214 WARREN & DAVIS

KARZINKIN G. S. & TARKOVSKAYA (1962) [Determination of caloric value ofsmall samples.], pp. 122-124. In Pavlovsky E. N. (Ed.) Techniques for the Investiga-tion of Fish Physiology, Akademiya Nauk SSSR Ikhtiologicheskaya Komissiya.(OTS 64-1101.)

KINNE 0. (196o) Growth, food intake and food conversion in an euryplastic fishexposed to different temperatures and salinities. Physiol. Zool., 33, 288-317.

KLEIBER M. (1926) Kaninchen und Ochsen als Futterrerwerter. Die Tierwelt, 36, 437.KLEIBER M. (1933) Tiergrosse und Futterverwertung. Tierernahrung, 5, 1-12.KLEIBER M. (1936) Problems involved in breeding for efficiency of food utilization.

Rec. Proc. Am. Soc. Anim. Prod., 27 Nov., 1938.KLEIBER M. (196i a) The Fire of Life, an Introduction to Animal Energetics. New York,

John Wiley & Sons, Inc. 454 pp.KLEIBER M. (196,b) Metabolic rate and food utilization as a function of body size.

Brody Memorial Lecture I. Res. Bull. Mo. Agric. Exp. Sta., 767,1-42.LINDEMAN R. L. (1942) The trophic dynamic aspect of ecology. Ecology, 23, 399-418.MArrLA/vo P. S. (1965) The feeding relationships of salmon, trout, minnows, stone

loach and three-spined sticklebacks in the River Endrick, Scotland. J. Anim.Ecol., 34, 109-133.

MANN K. H. (1965) Energy transformations by a population of fish in the RiverThames. J. Anim. Ecol., 34, 253-275.

NEILL R.M. (1938) The food and feeding of the brown trout (Salmo trutta L.) inrelation to the organic environment. Trans. R. Soc. Edinb., 59, 481-520.

PALOHEIMO J.E. & DICKIE L.M. (1965) Food and growth of fishes. I. A growthcurve derived from experimental data. J. Fish. Res. Bd Can., 22,521-542.

PALOHEIMO J.E. & DICKIE L.M. (1966) Food and growth of fishes. II. Effects of foodand temperature on the relation between metabolism and body weight. J. Fish.Res. Bd Can., 23, 869-908.

PARKER R. R. & LARKIN P.A. (1959) A concept of growth in fishes. J. Fish. Res. BdCan., 721-745.

PENTELOW F. T.K. (1939) The relationship between growth and food consumptionin brown trout (Salmo trutta). J. exp. Biol., i6, 446-473.

STEWART N.E. (1962) The Influence of Oxygen Concentration on the Growth ofJuvenile Largemouth Bass. M.S. Thesis, Oregon State University, Corvallis, 44 PP.

TERROINE E. F. & WURMSER R. (1922) L'energie de croissance. I. Le development deAspergillus niger. Bull. Soc. Chim. biol., 4, 519-567.

WARREN C. E., WALES J. H., DAVIS G. E. & DOUDOROFF P. (1964) Trout productionin an experimental stream enriched with sucrose. J. Wildl. Mgmt., 28, 617-660.

WINBERG G. G. (1956) [Rate of metabolism and food requirements of fishes]. NauchnyeTrudy Belorusskovo Gosudarstvennovo Universiteta im. V. I. Lenina, Minsk.253 pp. (Fish. Res. Bd Can. Transl. Ser. 194.)

laboratory studies on the feeding, bioenergetics, and...

Documents